Projects

S6GRAN: Sustainable 6G Radio Access Networks: Methods, Models, and Algorithms (Co-PI)

Funder: Research Council of Finland

When entering towards the 6G mobile communications era, the huge numbers of connected devices with increased data traffic and new services imply significant challenges for the involved RAN design to be able to provide the system capacity and QoS with sustainability. This project focuses on developing new scientific knowledge and novel wireless transceiver and system optimization solutions enabling sustainable 6G radio access networks. The core academic collaborators include Chalmers University of Technology, University of Luxembourg, and Queen's University of Belfast. Industrial collaboration is also conducted largely with several companies creating the 6G technology.

DYNAMICS: Distributed Multi-Modal Sensing Aided Large-Scale MIMO Communications (PI)

Funder: Research Council of Finland

Joint NSF - RCF project with Prof. Ahmed Alkhateeb at Arizona State University, USA.

Driven by the increasing demand for higher data rates, current and future communication networks rely heavily on large-scale MIMO systems and higher-frequency bands. This, however, pose critical challenges, including high channel acquisition overhead, sensitivity to blockages, and difficulties in user scheduling and association. Recently, there has been growing interest in leveraging multi-modal sensory data to enhance perception and situational awareness. In this project, we aim to develop frameworks that consider both the overheads and benefits of such networks, optimize environment semantic extraction and data fusion at distributed sensors, and enhance overall network performance. The scientific innovations of this project will directly contribute to enabling scalable and reliable operation of high-frequency MIMO communication systems.

DIRECTION: Deep Unfolding 6G Wireless Communications and Sensing Solutions (PI)

Funder: Research Council of Finland

To realize such a rapid growth of the data traffic and applications, wideband THz communications, reconfigurable intelligent surface (RIS), and joint communications and sensing (JCAS/ISAC) are emerging technologies enabling substantial improvements in both communications and sensing capabilities. However, their potentials are only fully unlocked with the development of low-complexity, scalable, and practical design solutions. DIRECTION aims at developing efficient deep unfolding solutions to 6G signal processing problems, with a focus on, but not limited to, the designs and optimization of wideband THz transceivers, metasurface controlling, and JCAS. The proposed solutions leverage the immense potential of fusing domain knowledge into AI/ML, which defines a comprehensive design methodology of AI signal processing techniques for 6G systems.

INTEGRATE: Integrated Sensing and Communication with Reconfigurable Intelligent Surface and Machine Learning (PI)

Funder: Nokia Donation

Joint project with Prof. Ping Jack Soh's antenna team at CWC-UOulu.

ISAC is anticipated to be a cornerstone technology in future wireless communication networks, enabling seamless coexistence of data transmission and environmental sensing. However, the dual-function operations in ISAC often lead to significant performance trade-offs. To address this challenge, RISs present a promising solution, as they can reshape the EM environment and enhance ISAC system performance. Coupling RIS with ML further allows for efficient operation with reduced complexity and faster execution, facilitating the practical deployment. This project focuses on the design, development, and demonstration of ML-enabled RIS-assisted ISAC solutions. The partnership will leverage expertise in algorithmic innovation and hardware implementation to explore and validate the potential of ISAC-RIS in the 6-15 GHz frequency range, a prime candidate for 6G deployments.

DECENT: Deep Unfolding Solutions to Energy-Efficient Transceiver Designs in 6G Wideband THz Systems (PI)

Funder: Nokia Foundation

DECENT aims at developing efficient artificial intelligence (AI)/machine learning (ML) solutions to hybrid beamforming (HBF) transceiver designs for terahertz (THz) communications. An efficient AI/DL solution here indicates a well-configured deep neural network (DNN) that not only achieves a high spectral or energy efficiency but also has a lightweight network architecture for scalable, distributed, and energy-efficient deployment. Our fundamental research goal is to provide a method to replace cumbersome iterative algorithms with efficient deep unfolding solutions to wideband transceiver design problems.

Page updated

Google Sites

Report abuse